atomicthumbs

They claim to have done that already, with little result. They don't care very much about this.

why do you think "lighting the world" is good in this case

is it cheaper to build a cellular tower or a satellite

You have never gone outside somewhere it's actually dark if you think this is the case. They're extremely visible after midnight.

this seems slightly delusional

do you believe everything Musk says

telling someone to read something written by a machine is disrespectful of their time. if you didn't care enough to write it, why should anyone else care enough to read it? its value is zero.

so poor people don't get to do to school?

This sort of attitude is how you get a culturally impoverished nation.

yes, they should

i don't think that has anything to do with this paper. isn't parallel tool calling just "assemble several tool calls at once"

5G NR radios could also be capable of communicating peer to peer long distance if manufacturers (phone or baseband?) wanted them to. It's in the spec.

New paper out of the Max Planck Institute for Intelligent Systems. If this holds up, it seems big.

Abstract: The continued improvements in language model capability have unlocked their widespread use as drivers of autonomous agents, for example in coding or computer use applications. However, the core of these systems has not changed much since early instruction-tuned models like ChatGPT. Even advanced AI agents function on message exchange formats, successively exchanging messages with users, systems, with itself (i.e. chain-of-thought) and tools in a single stream of computation. This bottleneck to a single stream in chat models leads to a number of limitations: the agent cannot act (generate output) while reading, and in reverse, cannot react to new information while writing. Similarly, the agent cannot act while thinking and cannot think while reading or acting on information. In this work, we show that models can be unblocked by switching from instruction-tuning for sequential message formats to instruction-tuning for multiple, parallel streams of computation, splitting each role into a separate stream. Every forward pass of the language model then simultaneously reads from multiple input streams and generates tokens in multiple output streams, all of which causally depend on earlier timesteps. We argue that this data-driven change remedies a number of usability limitations as outlined above, improves model efficiency through parallelization, improves model security through better separation of concerns and can further improve model monitorability.

"Hundreds of companies rely on Stainless to generate SDKs, CLIs, and MCP servers—the libraries, command-line tools, and connectors that let developers and agents use an API."

not anymore lol

is it because amateur radio operators legally have standards they have to comply with?

you'd think with how often Opus builds two separate code paths without feature parity when you try to vibe code something complex, people wouldn't regard this whole thing so highly

Yes, I know. I'm pointing out they're comparing a laboratory prototype to a commercial product.

And how much commercial development have NMC and LFP batteries had since they left the laboratory?

eBay Seller Research shows that the average sold price for a 5800X3D has increased by about $100 in the past 30 days, from ~$360 to ~$460.

not really, because all the sousveillance in the world doesn't grant the average joe the power of a single cop

There are more things in this world than software. Many of them are important!

100%

AIs think like a rock flies.

Virtually all successful existing sequence models rely on mean squared error (MSE) or dot-product similarity for both their bias and retention. This reliance can make models sensitive to outliers and limit their expressive power.

[...]

MEMORA: This model focuses on achieving the best possible memory stability by forcing its memory to act like a strict probability map. By using this constraint, it ensures that every time the memory state is updated, the changes are controlled and balanced. This guarantees a clean, stable process for integrating new information.Virtually all successful existing sequence models rely on mean squared error (MSE) or dot-product similarity for both their bias and retention. This reliance can make models sensitive to outliers and limit their expressive power.

so did a Titans write this

The person you're replying to asked for a source, not an anecdote.

Those aren't iMessage hardware backdoors.

Do you have a testable hypothesis about this or are you flailing in the dark?

these things are popping "ordinary" adults' minds like popcorn kernels and you want to take their safeguards off... why?

better not use UPS; goods moving under ATA carnet are on their prohibited list

"Im from a totally normal middle class-ish family."

Maybe that's why you can't comprehend the idea of being poor in a fashion that doesn't let you escape, even working as hard as you possibly can and making all the right choices.